PDA

View Full Version : using diff tracking


zulu
27th November 2001, 13:00
in my next ripp i want to give fairuse's new diff tracking method a try. How many encodings @ which diff would u suggest?

thanx in advance

pacohaas
27th November 2001, 16:58
depends on what you are focusing on, size or quality. Check my DT guide out.

joshcali
12th December 2001, 09:25
heya pacohaas
I looked at your diff track guide, it's a little
hard to understand.
Is it suggesting that you do 2 test encodings for
every rip to figure out diff values?

pacohaas
12th December 2001, 17:59
yes, I use method 3. Unless you feel lucky enough to guess the proper Diff value for your particular movie/resolution, this is the fastest.

joshcali
13th December 2001, 08:32
so your saying do 2 diff passes at 8000 and 12000, then see what their size is. use these sizes in the diff calc to get your average
diff value, then do 2 more encodings on either side of this diff value by 250. combine these 2.

is there no way to get a decent guess at a diff value? I'd love to just do 1 or 2 passes instead of the 4 that are used now...

pacohaas
13th December 2001, 20:21
There's no way to guess what diff value will give you what size for a particular movie. There's just too many variables. Based on a power relationship, it can be figured out, that's what my program does, but this requires 2 encodings. I like it because it allows me to extract the audio on the first pass(8000+12000), then encode a VBR mp3 and do the second pass accordingly. So, it's really only 2 passes, but 4 encodings (semantics really, i knew what you meant)